基于P-Contourlet的纹理图像检索
|
摘要
随着数字图像在人们生活中的快速增长,如何从海量的图像中搜寻特定的图像变为一个热点和难点。纹理是表征自然图像的一个重要特征,它能使人产生心理和视觉两方面的冲击。基于信号处理的纹理分析方法建立在时频分析与多分辨率分析的基础之上。心理物理学研究表明：人们在观察图像时,大脑会对图像进行频率分析。因此,基于信号处理的纹理分析方法是符合人类视觉与心理感受的。
本文主要研究基于信号处理的纹理分析方法来检索数字图像的问题。在对传统的多分辨率分析方法进行详细研究并进行大量纹理检索实验的前提下,提出了一种新的基于投影的Contourlet变换——P-Contourlet变换。该变换具有多分辨率、多方向选择性和良好的平移不变性,同时,包含丰富的相信息以及相对较低的冗余度。文章具体所做工作如下：
1.阐述了图像检索的背景与研究意义,介绍了该研究领域主要的研究成果及发展趋势,并对目前现有的基于纹理分析的图像检索技术做了简单的总结。
2.详细介绍了目前广为流行的小波变换、双树复小波变换、Contourlet变换、非子采样Contourlet变换、PDTDFB变换等方法,并将其应用到纹理图像检索中,进行了相关的实验分析。
3.提出了一个新的平移不变复Contourlet变换——P-Contourlet,详细介绍了该变换的结构及原理,阐述了它的平移不变性、多尺度性、多方向选择性、包含相信息、完备性以及较低冗余度等优点。并将其引入到纹理检索中,通过实验验证了其表征纹理特征的高效性。
4.提出了一种基于P-Contourlet变换和隐马尔科夫树(HMT)模型的纹理图像检索方法。该方法用隐马尔科夫树模型对P-Contourlet各子带系数进行建模,将其作为图像的纹理特征,实验验证了该模型对P-Contourlet各子带系数拟合的有效性以及用该方法进行图像检索的高效性。
With the rapid growth of the digital image in people's lives, how to search for a specific image from a large of images dataset becomes a hot and difficult question. Texture is an important feature of natural images, it can engender both psychological and visual impact. Signal processing based texture analysis method is established on the time-frequency analysis and multiresolution analysis. Psychophysical studies show that the brain will analysis the frequency components of images when people observe it. So signal processing based texture analysis method is fit human's visual and psychological feelings.
This paper mainly studies on signal processing based texture image retrieval. Under the premise of large number experiments and detailed studies on traditional multi-resolution analysis based texture retrieval, we propose a new projection-based contourlet transform-P-Contourlet transform. Which is multi-resolution, multi-directional selectivity, nearly shift invariance, rich phase information and low redundancy. The main work is as follows:
1. Describes the background and research significance of image retrieval. Review some major research results and development trends in this filed. Make a brief summary about texture based image retrieval methods in presently.
2. Makes a detail studies on wavelet transform, dual-tree complex wavelet transform, Contourlet transform, Nonsubsampled Contourlet Transform and PDTDFB transform, which was widely popular in present. Apply these transforms to texture image retrieval and do a large number experimental analysis.
3. Presents a new shift invariant complex contourlet transform—P-Contourlet transform. Which is multi-resolution, multi-directional selectivity, nearly shift invariance, rich phase information and low redundancy. Apply it to texture image retrieval and the experiments shows it's efficient.
4. Presents a new texture image retrieval method based on P-Contourlet transform and Hidden Markov Tree (HMT) model. By modeling the P-Contourlet transform's each subbands coefficients using hidden Markov tree, obtains the images texture feature vectors. The experiments shows that HMT model is validity in modeling P-Contourlet transform's each subbands coefficients and this retrieval method is efficiency.
本文主要研究基于信号处理的纹理分析方法来检索数字图像的问题。在对传统的多分辨率分析方法进行详细研究并进行大量纹理检索实验的前提下,提出了一种新的基于投影的Contourlet变换——P-Contourlet变换。该变换具有多分辨率、多方向选择性和良好的平移不变性,同时,包含丰富的相信息以及相对较低的冗余度。文章具体所做工作如下：
1.阐述了图像检索的背景与研究意义,介绍了该研究领域主要的研究成果及发展趋势,并对目前现有的基于纹理分析的图像检索技术做了简单的总结。
2.详细介绍了目前广为流行的小波变换、双树复小波变换、Contourlet变换、非子采样Contourlet变换、PDTDFB变换等方法,并将其应用到纹理图像检索中,进行了相关的实验分析。
3.提出了一个新的平移不变复Contourlet变换——P-Contourlet,详细介绍了该变换的结构及原理,阐述了它的平移不变性、多尺度性、多方向选择性、包含相信息、完备性以及较低冗余度等优点。并将其引入到纹理检索中,通过实验验证了其表征纹理特征的高效性。
4.提出了一种基于P-Contourlet变换和隐马尔科夫树(HMT)模型的纹理图像检索方法。该方法用隐马尔科夫树模型对P-Contourlet各子带系数进行建模,将其作为图像的纹理特征,实验验证了该模型对P-Contourlet各子带系数拟合的有效性以及用该方法进行图像检索的高效性。
With the rapid growth of the digital image in people's lives, how to search for a specific image from a large of images dataset becomes a hot and difficult question. Texture is an important feature of natural images, it can engender both psychological and visual impact. Signal processing based texture analysis method is established on the time-frequency analysis and multiresolution analysis. Psychophysical studies show that the brain will analysis the frequency components of images when people observe it. So signal processing based texture analysis method is fit human's visual and psychological feelings.
This paper mainly studies on signal processing based texture image retrieval. Under the premise of large number experiments and detailed studies on traditional multi-resolution analysis based texture retrieval, we propose a new projection-based contourlet transform-P-Contourlet transform. Which is multi-resolution, multi-directional selectivity, nearly shift invariance, rich phase information and low redundancy. The main work is as follows:
1. Describes the background and research significance of image retrieval. Review some major research results and development trends in this filed. Make a brief summary about texture based image retrieval methods in presently.
2. Makes a detail studies on wavelet transform, dual-tree complex wavelet transform, Contourlet transform, Nonsubsampled Contourlet Transform and PDTDFB transform, which was widely popular in present. Apply these transforms to texture image retrieval and do a large number experimental analysis.
3. Presents a new shift invariant complex contourlet transform—P-Contourlet transform. Which is multi-resolution, multi-directional selectivity, nearly shift invariance, rich phase information and low redundancy. Apply it to texture image retrieval and the experiments shows it's efficient.
4. Presents a new texture image retrieval method based on P-Contourlet transform and Hidden Markov Tree (HMT) model. By modeling the P-Contourlet transform's each subbands coefficients using hidden Markov tree, obtains the images texture feature vectors. The experiments shows that HMT model is validity in modeling P-Contourlet transform's each subbands coefficients and this retrieval method is efficiency.
引文
[1]G. Salton and C. Buckley. Term weighting approaches in automatic text retrieval. Information Processing and Management,24(5):pp.513-523,1988.
[2]Enser. Pictorial information retrieval. Journal of Documentation,51(2), pp.126-170.2002.
[3]Aslandogan Y A, Yu C T. Techniques and Systems for Image and Video Retrieval. IEEE Trans on Knowledge and Data Engineering, Vol.11, No.1, pp.56-60.1999.
[4]刘洋.基于内容的图像检索算法研究[D].吉林大学,2005.
[5]张李秋.一种基于纹理特征提取的图像检索方法[D],电子科技大学,2007.
[6]Kun Xu, Peter Meer, etal. Performance Analysis in Content-based Retrieval with Textures. In Proceedings.15th International Conference on Pattern Recognition, pp.275-278,2000.4.
[7]Haralick, Shanmugnm, Dinstein. Textural features for image classification[J]. IEEE Transactions on Systems, Man. And Cybernetics,3(6):pp.610-621,1973
[8]Haralick. Statistical and structural approaches to texture [J]. Proceedings of the IEEE,67(5): pp.786-804,1979.
[9]Tuceryan, Jain. Texture Analysis, Handbook Pattern Recognition and Computer Vision [M]. Singapore:World Scientific. pp.235-276,1993.
[10]刘丽,匡纲要.图像纹理特征提取方法综述[J].中国图象图形学报.pp.622-635.2009.4
[11]Kingsbury N G. Complex wavelets for shift invariant analysis and filtering of signals [J]. Appl. Comp. Harmonic Anal,10 (3):pp.234-253.2000.
[12]Kingsbury N G. The dual-tree complex wavelet transform:A new efficient tool for image restoration and enhancement [A].Proc. European Signal Processing Conf.[C]. pp.319-322. 1998
[13]R. H. Bamberger, M. J. T. Smith. A filter bank for the directional decomposition of images: theory and design [J]. IEEE Transactions on Signal Processing,40(4):pp.882-893.1992.
[14]Arthur L. da Cunha, Jianping Zhou, Minh N. Do. The nonsubsampled contourlet transform: theory, design and applications. IEEE Trans. Image Processing, Vol.15, No.10, pp. 3089-3101, Oct.2006.
[15]Truong T. Nguyen, Soontorn Oraintara. The shiftable complex directional Pyramid-Part I: theoretical aspects. IEEE Trans. Signal Processing, Vol.56, No.10, pp.4651-4660, Oct.2008.
[16]Truong T. Nguyen, Soontorn Oraintara. The shiftable complex directional Pyramid-Part II: implementation and applications. IEEE Trans. Signal Processing, Vol.56, No.10, pp. 4661-4672, Oct.2008.
[17]Brodatz album, http://sipi.usc.edu/database/database.cgi?volume=textures&image=12
[18]曲怀敬Contourlet变换在纹理图像检索和医学图像分割中的应用研究[D].山东大学,2009.
[19]R. L. Joshi, T. R. Fisher. Comparison of generalized Gaussian and Laplacian modeling in DCT Image coding [J].IEEE. Signal Processing Letters.2(5):pp.81-82.1995.
[20]P. Moulin, J. Liu. Analysis of multiresolution image denoising scheme using generalized Gaussian and complexity priors [J]. IEEE Transactions on Information Theory.45(3): pp.909-919.1999.
[21]Minh N. Do and Martin Vetterli. Wavelet-based texture retrieval using generalized Gaussian density and kullback-leibler distance. IEEE Transactions on Image Processing, Vol.11, pp. 146-158, February 2002.
[22]T. M. Cover, J. A. Thomas. Elements of information Theory[M].New York:Wiley,1991.
[23]S. Mallat. A Wavelet Tour of Signal Processing.2nd ed. New York:Academic,1999.
[24]D. L. Donoho, M. Vetterli, R. A. DeVore, and I. Daubechies, Data compression and harmonic analysis. IEEE Trans. Inf. Theory, vol.44, no.6, pp.2435-2476, Oct.1998.
[25]M. N. Do and M. Vetterli. The contourlet transform:An efficient directional multiresolution image representation. IEEE Trans. Image Process. vol.14, pp.2107-2116, Dec.2005.
[26]F. C. A. Fernandes, R. van Spaendonck, M. Coates, C. S. Burrus. Directional complex-wavelet processing. Wavelet Applications Ⅶ, Proceedings of SPIE, San Diego, USA. pp.1-17,2000.
[27]张久文.方向性多分辨率图像分析研究：理论和应用[D].兰州大学,2009.
[28]Jiuwen Zhang, Jincai Mi, Tongfeng Zhang. Shift-invariant Texture Retrieval using P-Contourlet. International Conference on Advancing in Computing, Communications and Informatics (ICACCI-2012). pp.879-883,2012.
[29]D. Gabor, Theory of Communication. IEEE Trans. Information Theory, Vol.9, No.3, pp. 76-87,1946.
[30]T. Bulow, GSommer. A novel approach to the 2D analytic signal. In F. Solina and A. Leonardis (Eds.), CAIP'99, Ljubljana, Slovenia, pp.25-32,1999.
[31]I. W. Selesnick. Lowpass filters realizable as all-pass sums:design via a new flat delay filter. IEEE Trans. Circuits Systems Part Ⅱ 46 (1) (January 1999) pp.40-50.
[32]H. W. Schuessler, P. Stepen. Recursive halfband-filters. Internat. J. Electron. Commun.55 (6) (2001) pp.1-12.
[33]E. P. Simoncelli, W. T. Freeman, E. H. Adelson and D. J. Heeger. Shiftable multiscale transforms. IEEE Trans. Inform. Theory,38(2):pp.587-607, March 1992.
[34]M. S. Crouse, R. D. Nowak and R. G. Baraniuk. Wavelet-Based Statistical Signal Processing using Hidden Markov Models. IEEE Trans Signal Proc, vol.46, No.4, pp.886-902, April 1998.
[35]H. Choi, and R.G. baraniuk. Multiscale Image Segmentation Using Wavelet Domain Hidden Markov Models. IEEE Trans. Image pro, vol.10, no.9, pp.1309-1321, Sep.2001.
[36]B. S. Raghavendra and P. Subbanna Bhat. On Hidden Markov Model-Contourlet Hidden markov Tree based Texture Segmentation. IEEE conferences signal processing. pp.126-130, April 2004.
[37]L. R. Rabiner, A Tutorial on Hidden Markov Models and Selected Applications in Speech recognition, Proc. IEEE,77(2):pp.257-286,1989
[38]M. N. Do. Fast Approximation of Kullback-Leibler Distance for Dependence Trees and Hidden Markov Models. IEEE Signal Processing Letters, pp.115-118. April,2003.
[39]M. N. Do. Rotation Invariant Texture Characterization and Retrieval using Steerable Wavelet-domain Hidden Markov Models. IEEE Trans. Multimedia, pp.1-12. December, 2002.
[2]Enser. Pictorial information retrieval. Journal of Documentation,51(2), pp.126-170.2002.
[3]Aslandogan Y A, Yu C T. Techniques and Systems for Image and Video Retrieval. IEEE Trans on Knowledge and Data Engineering, Vol.11, No.1, pp.56-60.1999.
[4]刘洋.基于内容的图像检索算法研究[D].吉林大学,2005.
[5]张李秋.一种基于纹理特征提取的图像检索方法[D],电子科技大学,2007.
[6]Kun Xu, Peter Meer, etal. Performance Analysis in Content-based Retrieval with Textures. In Proceedings.15th International Conference on Pattern Recognition, pp.275-278,2000.4.
[7]Haralick, Shanmugnm, Dinstein. Textural features for image classification[J]. IEEE Transactions on Systems, Man. And Cybernetics,3(6):pp.610-621,1973
[8]Haralick. Statistical and structural approaches to texture [J]. Proceedings of the IEEE,67(5): pp.786-804,1979.
[9]Tuceryan, Jain. Texture Analysis, Handbook Pattern Recognition and Computer Vision [M]. Singapore:World Scientific. pp.235-276,1993.
[10]刘丽,匡纲要.图像纹理特征提取方法综述[J].中国图象图形学报.pp.622-635.2009.4
[11]Kingsbury N G. Complex wavelets for shift invariant analysis and filtering of signals [J]. Appl. Comp. Harmonic Anal,10 (3):pp.234-253.2000.
[12]Kingsbury N G. The dual-tree complex wavelet transform:A new efficient tool for image restoration and enhancement [A].Proc. European Signal Processing Conf.[C]. pp.319-322. 1998
[13]R. H. Bamberger, M. J. T. Smith. A filter bank for the directional decomposition of images: theory and design [J]. IEEE Transactions on Signal Processing,40(4):pp.882-893.1992.
[14]Arthur L. da Cunha, Jianping Zhou, Minh N. Do. The nonsubsampled contourlet transform: theory, design and applications. IEEE Trans. Image Processing, Vol.15, No.10, pp. 3089-3101, Oct.2006.
[15]Truong T. Nguyen, Soontorn Oraintara. The shiftable complex directional Pyramid-Part I: theoretical aspects. IEEE Trans. Signal Processing, Vol.56, No.10, pp.4651-4660, Oct.2008.
[16]Truong T. Nguyen, Soontorn Oraintara. The shiftable complex directional Pyramid-Part II: implementation and applications. IEEE Trans. Signal Processing, Vol.56, No.10, pp. 4661-4672, Oct.2008.
[17]Brodatz album, http://sipi.usc.edu/database/database.cgi?volume=textures&image=12
[18]曲怀敬Contourlet变换在纹理图像检索和医学图像分割中的应用研究[D].山东大学,2009.
[19]R. L. Joshi, T. R. Fisher. Comparison of generalized Gaussian and Laplacian modeling in DCT Image coding [J].IEEE. Signal Processing Letters.2(5):pp.81-82.1995.
[20]P. Moulin, J. Liu. Analysis of multiresolution image denoising scheme using generalized Gaussian and complexity priors [J]. IEEE Transactions on Information Theory.45(3): pp.909-919.1999.
[21]Minh N. Do and Martin Vetterli. Wavelet-based texture retrieval using generalized Gaussian density and kullback-leibler distance. IEEE Transactions on Image Processing, Vol.11, pp. 146-158, February 2002.
[22]T. M. Cover, J. A. Thomas. Elements of information Theory[M].New York:Wiley,1991.
[23]S. Mallat. A Wavelet Tour of Signal Processing.2nd ed. New York:Academic,1999.
[24]D. L. Donoho, M. Vetterli, R. A. DeVore, and I. Daubechies, Data compression and harmonic analysis. IEEE Trans. Inf. Theory, vol.44, no.6, pp.2435-2476, Oct.1998.
[25]M. N. Do and M. Vetterli. The contourlet transform:An efficient directional multiresolution image representation. IEEE Trans. Image Process. vol.14, pp.2107-2116, Dec.2005.
[26]F. C. A. Fernandes, R. van Spaendonck, M. Coates, C. S. Burrus. Directional complex-wavelet processing. Wavelet Applications Ⅶ, Proceedings of SPIE, San Diego, USA. pp.1-17,2000.
[27]张久文.方向性多分辨率图像分析研究：理论和应用[D].兰州大学,2009.
[28]Jiuwen Zhang, Jincai Mi, Tongfeng Zhang. Shift-invariant Texture Retrieval using P-Contourlet. International Conference on Advancing in Computing, Communications and Informatics (ICACCI-2012). pp.879-883,2012.
[29]D. Gabor, Theory of Communication. IEEE Trans. Information Theory, Vol.9, No.3, pp. 76-87,1946.
[30]T. Bulow, GSommer. A novel approach to the 2D analytic signal. In F. Solina and A. Leonardis (Eds.), CAIP'99, Ljubljana, Slovenia, pp.25-32,1999.
[31]I. W. Selesnick. Lowpass filters realizable as all-pass sums:design via a new flat delay filter. IEEE Trans. Circuits Systems Part Ⅱ 46 (1) (January 1999) pp.40-50.
[32]H. W. Schuessler, P. Stepen. Recursive halfband-filters. Internat. J. Electron. Commun.55 (6) (2001) pp.1-12.
[33]E. P. Simoncelli, W. T. Freeman, E. H. Adelson and D. J. Heeger. Shiftable multiscale transforms. IEEE Trans. Inform. Theory,38(2):pp.587-607, March 1992.
[34]M. S. Crouse, R. D. Nowak and R. G. Baraniuk. Wavelet-Based Statistical Signal Processing using Hidden Markov Models. IEEE Trans Signal Proc, vol.46, No.4, pp.886-902, April 1998.
[35]H. Choi, and R.G. baraniuk. Multiscale Image Segmentation Using Wavelet Domain Hidden Markov Models. IEEE Trans. Image pro, vol.10, no.9, pp.1309-1321, Sep.2001.
[36]B. S. Raghavendra and P. Subbanna Bhat. On Hidden Markov Model-Contourlet Hidden markov Tree based Texture Segmentation. IEEE conferences signal processing. pp.126-130, April 2004.
[37]L. R. Rabiner, A Tutorial on Hidden Markov Models and Selected Applications in Speech recognition, Proc. IEEE,77(2):pp.257-286,1989
[38]M. N. Do. Fast Approximation of Kullback-Leibler Distance for Dependence Trees and Hidden Markov Models. IEEE Signal Processing Letters, pp.115-118. April,2003.
[39]M. N. Do. Rotation Invariant Texture Characterization and Retrieval using Steerable Wavelet-domain Hidden Markov Models. IEEE Trans. Multimedia, pp.1-12. December, 2002.